Papers › Leveraging Future Relationship Reasoning for Vehicle Trajectory Prediction

Leveraging Future Relationship Reasoning for Vehicle Trajectory Prediction

24 May 2023arXiv:2305.14715archive 2025-07-28

Daehee Park, Hobin Ryu, Yunseo Yang, Jegyeong Cho, Jiwon Kim, Kuk-Jin Yoon

Understanding the interaction between multiple agents is crucial for realistic vehicle trajectory prediction. Existing methods have attempted to infer the interaction from the observed past trajectories of agents using pooling, attention, or graph-based methods, which rely on a deterministic approach. However, these methods can fail under complex road structures, as they cannot predict various interactions that may occur in the future. In this paper, we propose a novel approach that uses lane information to predict a stochastic future relationship among agents. To obtain a coarse future motion of agents, our method first predicts the probability of lane-level waypoint occupancy of vehicles. We then utilize the temporal probability of passing adjacent lanes for each agent pair, assuming that agents passing adjacent lanes will highly interact. We also model the interaction using a probabilistic distribution, which allows for multiple possible future interactions. The distribution is learned from the posterior distribution of interaction obtained from ground truth future trajectories. We validate our method on popular trajectory prediction datasets: nuScenes and Argoverse. The results show that the proposed method brings remarkable performance gain in prediction accuracy, and achieves state-of-the-art performance in long-term prediction benchmark dataset.

PaperPDF

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Motion ForecastingPredictionTrajectory Prediction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Motion Forecasting Argoverse CVPR 2020 FRM DAC (K=6) 0.9878 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM MR (K=1) 0.5728 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM MR (K=6) 0.143 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM brier-minFDE (K=6) 1.9365 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM minADE (K=1) 1.7063 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM minADE (K=6) 0.8165 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM minFDE (K=1) 3.7486 #56 of 299 Archive leaderboard report
Motion Forecasting Argoverse CVPR 2020 FRM minFDE (K=6) 1.2671 #56 of 299 Archive leaderboard report
Trajectory Prediction nuScenes FRM MinADE_10 0.88 #4 of 34 Archive leaderboard report
Trajectory Prediction nuScenes FRM MinADE_5 1.18 #4 of 34 Archive leaderboard report
Trajectory Prediction nuScenes FRM MinFDE_1 6.59 #4 of 34 Archive leaderboard report
Trajectory Prediction nuScenes FRM MissRateTopK_2_10 0.30 #4 of 34 Archive leaderboard report
Trajectory Prediction nuScenes FRM MissRateTopK_2_5 0.48 #4 of 34 Archive leaderboard report
Trajectory Prediction nuScenes FRM OffRoadRate 0.02 #4 of 34 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

fail

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections